[garrytan]

In plain words

Since the 2026-10-15 recording (#3212), 54 new free test files have landed. CI therefore guesses their run time as the 99th-percentile duration, and the weekly health check fails CEO-10 again (limit 5). This PR re-records every free file's duration on a CI-mirrored Ubicloud VM, so the shard packer has real numbers for all of them. Only data files change.

What changed

• scripts/free-test-durations.json: recorded with UBI_OWNER=gsta33 bun run test:ubicloud --record-durations (standard-16) on origin/main cadbd0af (v1.91.116.0). The recorder exited 0 after 316 s, and no file failed or timed out. Durations went from 1219 to 1273: 54 added, none removed. The serial sum went from 3202 s to 3515 s.
• scripts/free-test-seed-allowlist.json: removes test/setup-codex-scope.test.ts, which now records 58.9 s, under the 60 s ceiling. The shrink-only ratchet requires dropping entries that no longer exceed it. The two other files over 60 s from v1.91.105.0 fix(tests): the free suite stops needing its retry pass; test:health tells the truth (#3154) #3212 remain allowlisted with their split todos: setup-install-registry at 86.5 s and host-renders-prune at 63.5 s.

free-tests.yml is untouched; #3236 owns its curation. A standing recorder is planned as a separate lane (W4).

Validation

• unseededCheck (the CEO-10 check in scripts/test-health-report.ts) on this tree: pass 0 unseeded file(s) (limit 5). On main it reports fail with 54 unseeded files.
• GSTACK_FREE_SEED_BASE=$(git merge-base HEAD origin/main) bun test test/free-seed-ratchet.test.ts test/test-health-report.test.ts: 26 pass, 0 fail.
• The diff is data only: no code, tests or prompts change, so no paid eval selection is expected.

Behavior changes

None. The shard packer reads the new durations.

Refs #3224. Expect CEO-10 to go red again as new test files accumulate until the standing recorder lands.

[trunk-io]

😎 This pull request was merged.

[github-actions]

E2E Evals: ✅ PASS

[test:paid] VERDICT GREEN — lane gate/pr, attempt 1
  rule 0/0 · behavior 0/0 · judge 0/0 · quarantined 0 (0 failing the lane)
  SKIPPED 0 · INFRA 0 · INCOMPLETE 0 · unattributed 0 · ACTION REQUIRED 0
  wall 0m00s · cost $0.00

0 executed, 0 reused rule/judge records | 0 manual accepted (unscored; no score-cache credit) | $0 rule/judge cost | reconcile exit: 0

Profile: pr / mode pr; selected behaviors: 0, judges: 0. Deferred to scheduled/release coverage: 0 behaviors and 0 changed prompt files. Deferred checks did not run and receive no PR-pass credit.

Rule and judge shards

Shard	Automated result	Manual/unscored	Executed	Reused	Status	Cost

Fail-closed reconciliation

[test:paid] report: 1/1 slices, 0 planned shards, tier=gate, attempt 1
[test:paid] coverage: profile=pr mode=pr; selected E2E=0, judges=0
[test:paid] deferred: 0 broad behaviors, 0 changed prompts without quick live coverage; these are not PR passes
[test:paid] collector final outcomes: 0 executed, 0 reused; 0 passed, 0 failed, 0 manual accepted (unscored; no score-cache credit) (0 attempt records from 0 collectors; every record counts)
[test:paid] VERDICT GREEN — lane gate/pr, attempt 1
  rule 0/0 · behavior 0/0 · judge 0/0 · quarantined 0 (0 failing the lane)
  SKIPPED 0 · INFRA 0 · INCOMPLETE 0 · unattributed 0 · ACTION REQUIRED 0
  wall 0m00s · cost $0.00
[test:paid] report: every planned shard accounted and passed

---

Sliced lane: planner → duration-packed executors → fail-closed report. Behavior cases run a pre-registered 3-trial panel (PASS at 2/3 with no contract violation); a PASS 2/3 is shown with its failed trial, never as a clean pass. Reused results retain their original provenance and expiry.